# AI transparency statement URL: https://vrisic.com/ai-transparency/ Last updated: 2026-10-04 > Where we use AI on this site, how we tell you, and the rules we follow when we build AI for clients. **Last updated:** 4 October 2026 On this page 1. AI on this website 2. How we tell you 3. What our assistants can and cannot do 4. Your data and the AI providers 5. AI generated content 6. How we build AI for clients 7. Rules we follow 8. Questions and concerns ## 1. AI on this website - **Chat assistant.** The chat button opens an AI assistant that answers questions about our services, prices and process, and can pass your details to our team. - **Voice advisor.** The voice call button starts a spoken conversation with an AI advisor. If we switch on our phone line, the same advisor may answer calls to our phone number. - **Explainer videos.** The narration in the short videos on our pages is a synthetic AI voice reading a script our team wrote. - **Illustrations.** Diagrams and illustrations on the site are drawn by our team with code. They show example data, not real client results. ## 2. How we tell you Our assistants say they are AI at the start of each conversation and whenever you ask. You never need to guess whether you are talking to a person. This follows Article 50 of the EU Artificial Intelligence Act, which has required this kind of disclosure since 2 August 2026, and similar rules and guidance in other countries. You can ask for a person at any time. The chat moves you to WhatsApp with our team, and the voice advisor arranges a call back from a named person on our team. ## 3. What our assistants can and cannot do - They share general information and typical price ranges from our published facts. A binding price only comes in a written proposal. - They do not make decisions about you. Every enquiry is reviewed by a person, and no AI output on this site has a legal or similarly significant effect on you. - They can make mistakes. If an answer matters to you, please confirm it with our team. - They are instructed not to ask for passwords, card numbers, health details or other sensitive data. Please do not share them. - They check that an email address or phone number looks valid before saving it, and ask you to type it if speech was unclear. ## 4. Your data and the AI providers To answer you, the assistants send your messages, or the text of what you said, to AI model providers listed on our [subprocessors](https://vrisic.com/subprocessors/) page. Voice calls in Chrome, Edge and Safari use your browser’s own speech recognition, which may send audio to your browser’s maker. In other browsers we record your answer and transcribe it on our server. We keep the text of conversations, not audio recordings. How long we keep it and your rights are in our [privacy policy](https://vrisic.com/privacy-policy/) . Some free AI plans allow the provider to use content to improve its services. That is one more reason not to share sensitive data with any website chat, including ours. Client projects always use paid or private model access where the provider does not train on your data. ## 5. AI generated content We write and review the articles on this site ourselves. When we use AI to help draft, translate or check text, a person on our team edits and approves it before it is published. Synthetic voices in our videos are labelled on this page and never imitate a real person. ## 6. How we build AI for clients - **Risk first.** At scoping we classify each use case under the EU AI Act categories (prohibited, high risk, limited risk, minimal risk) and the rules of your country, and we tell you what that means for the build. - **Human oversight.** Systems that affect people, money, health or legal matters keep a person in the loop, with clear review and override steps. - **Testing.** We measure accuracy on an evaluation set built from your data before launch, and keep testing after launch. - **Disclosure built in.** Chat and voice agents we build tell users they are AI, and we can add machine readable labels to generated content where the law requires it. - **Documentation.** We document intended use, data sources, known limits and how to oversee the system, so you can meet your own duties as the deployer. - **Your data stays yours.** Systems run in your cloud or a region we agree, and we do not use your data to train models for anyone else. We will not build systems on the prohibited list in our [acceptable use policy](https://vrisic.com/acceptable-use/) . ## 7. Rules we follow - **European Union:** the EU AI Act, including its transparency duties, and the GDPR, including the limits on decisions made only by automated means. - **United Kingdom:** UK GDPR, the Data (Use and Access) Act 2025 and guidance from the Information Commissioner’s Office on AI and data protection. - **United States:** state privacy laws such as the CCPA and CPRA, rules on automated decision making technology, and new state AI laws such as Colorado’s, which is due to apply from 1 January 2027. For hiring tools used in New York City we support the bias audit and notice rules. - **United Arab Emirates and Saudi Arabia:** the UAE Personal Data Protection Law and the UAE’s published AI ethics principles, and Saudi Arabia’s Personal Data Protection Law and the AI Ethics Principles from SDAIA. - **India:** the Digital Personal Data Protection Act, 2023 and its Rules, and the IT Rules on synthetically generated information. Laws on AI are changing quickly. We review this statement when they change. ## 8. Questions and concerns If you think one of our assistants got something wrong, treated you unfairly or did not tell you it was AI, tell us and a person will look into it: - Email: [info@vrisic.com](mailto:info@vrisic.com) - WhatsApp: [+91 63777 67206](https://wa.me/916377767206?text=Hi%20Vrisic%2C%20I%20found%20you%20on%20your%20website%20%28Privacy%20or%20legal%20question%29.%20I%20would%20like%20to%20talk%20about%20an%20AI%20project.) - Postal address: India (remote first company) --- Vrisic · https://vrisic.com/ · Contact: info@vrisic.com · Book a call: https://vrisic.com/contact/